Pain research relies heavily on traumatic models of nerve injury such as SNI. While these model well some aspects of neuropathic pain, the majority of neuropathy seen in the clinic is non-traumatic (e.g. diabetic, chemotherapy-induced, toxic, genetic, etc.)
In your experience, what has been the best performing neuropathy model? What do I mean by ‘best performing’?
The manipulation produces easily detected pain-like behaviors in the majority of animals
This is probably the most important aspect for me → Are there discernible and long-lasting behavioral effects?
SNI, in my hands, produces mechanical hypersensitivity in nearly every animal and it is easy to detect using Von Frey or cotton brush.
In my hands, non-traumatic models such as paclitaxel-induced neuropathy, do not show as strong behavioral effects and there is more variability amongst cohorts. I have heard this from other labs too. Some labs see strong hypersensitivity, others not. Some with just some assays, but only for a time.
Are there any non-traumatic neuropathy models that are as reliable in producing pain-like behaviors, analogous to SNI?
There is histological evidence of sensory neuron loss/degeneration
Obvious loss in PGP9.5 fibers in skin
The model uses a clinically-relevant manipulation: chemotherapy, diabetes, something else?
I don’t have anything to add, but want to say this is my experience as well. SNI actually works TOO well, in that it produces basically maximal allodynia in basically every mouse, which of course is very much unlike the clinical situation, in which nerve injuries only produce pain sometimes. I published a paper on paclitaxel allodynia back in the day (Smith et al., 2004) where we compared 10 strains. It worked pretty modestly in all of them, frankly, especially so in C57BL/6.
I have the same impression about the models. SNI produces robust mechanical allodynia and significative cold allodynia in acetone test. This is a problem when you are interesting in endogenous inhibitory mechanisms. To overcome, we normally use the von Frey filament in frequency model besides the up and down. So we can observe enhanced hypersensitivity in some cases (eg KO for some antinociceptive gene). In rare cases autotomy.
Regarding paclitaxel, we performed a dose response curve (2, 4 and 8 mg/kg/day four injections). We found that only the dose of 8 mg/kg produced a consistent mechanical and cold allodynia. On the other hand, this dose was the only one that produced hypoalgesia in Hargreaves test . We also observed a robust sensory neuronal loss/retraction in the paw skin (We used Nav1.8Cre Tdtomato animals). These phenotypes are long-lasting (up to 35 days).
Noteworthy, all these data were in B6 mice.
Also, it depends on what you mean by “model”. I’m trying to get people to stop using that word, since it’s never clear whether you’re referring to subject, assay, or measure.